Cover Image

A hybrid differential evolution algorithm for parameter tuning of evolving spiking neural network

In this paper, differential evolution (DE) has been utilised to solve the problem of tuning the parameters of evolving spiking neural network (ESNN) manually. As ESNN is sensitive to its parameters as other models, optimal integration of parameters leads to better classification accuracy. A hybrid d...

Full description

Saved in:
Bibliographic Details
Main Authors: Saleh, Abdulrazak Yahya, Shamsuddin, Siti Mariyam, Abdull Hamed, Haza Nuzly
Format: Article
Published: Inderscience Publishers 2017
Subjects:
QA75 Electronic computers. Computer science
Online Access:http://eprints.utm.my/id/eprint/97057/
http://dx.doi.org/10.1504/IJCVR.2017.081231
Tags: Add Tag
No Tags, Be the first to tag this record!
id my.utm.97057
record_format eprints
spelling my.utm.970572022-09-15T04:44:57Z http://eprints.utm.my/id/eprint/97057/ A hybrid differential evolution algorithm for parameter tuning of evolving spiking neural network Saleh, Abdulrazak Yahya Shamsuddin, Siti Mariyam Abdull Hamed, Haza Nuzly QA75 Electronic computers. Computer science In this paper, differential evolution (DE) has been utilised to solve the problem of tuning the parameters of evolving spiking neural network (ESNN) manually. As ESNN is sensitive to its parameters as other models, optimal integration of parameters leads to better classification accuracy. A hybrid differential evolution for parameter tuning of evolving spiking neural network (DEPT-ESNN) is presented for parameter optimisation for determining the optimal number of evolving spiking neural network (ESNN) parameters: modulation factor (Mod), similarity factor (Sim) and threshold factor (C). The best values of parameters are adaptively selected by differential evolution (DE) to avoid selecting suitable values for a particular problem by trial-and-error approach. Several standard datasets from UCI machine learning are used for evaluating the performance of this hybrid model. It has been found that the classification accuracy and other performance measures can be increased by using hybrid method with differential evolution DEPT-ESNN. Inderscience Publishers 2017 Article PeerReviewed Saleh, Abdulrazak Yahya and Shamsuddin, Siti Mariyam and Abdull Hamed, Haza Nuzly (2017) A hybrid differential evolution algorithm for parameter tuning of evolving spiking neural network. International Journal of Computational Vision and Robotics, 7 (1-2). pp. 20-34. ISSN 1752-9131 http://dx.doi.org/10.1504/IJCVR.2017.081231 DOI : 10.1504/IJCVR.2017.081231
institution Universiti Teknologi Malaysia
building UTM Library
collection Institutional Repository
continent Asia
country Malaysia
content_provider Universiti Teknologi Malaysia
content_source UTM Institutional Repository
url_provider http://eprints.utm.my/
topic QA75 Electronic computers. Computer science
spellingShingle QA75 Electronic computers. Computer science
Saleh, Abdulrazak Yahya
Shamsuddin, Siti Mariyam
Abdull Hamed, Haza Nuzly
A hybrid differential evolution algorithm for parameter tuning of evolving spiking neural network
description In this paper, differential evolution (DE) has been utilised to solve the problem of tuning the parameters of evolving spiking neural network (ESNN) manually. As ESNN is sensitive to its parameters as other models, optimal integration of parameters leads to better classification accuracy. A hybrid differential evolution for parameter tuning of evolving spiking neural network (DEPT-ESNN) is presented for parameter optimisation for determining the optimal number of evolving spiking neural network (ESNN) parameters: modulation factor (Mod), similarity factor (Sim) and threshold factor (C). The best values of parameters are adaptively selected by differential evolution (DE) to avoid selecting suitable values for a particular problem by trial-and-error approach. Several standard datasets from UCI machine learning are used for evaluating the performance of this hybrid model. It has been found that the classification accuracy and other performance measures can be increased by using hybrid method with differential evolution DEPT-ESNN.
format Article
author Saleh, Abdulrazak Yahya
Shamsuddin, Siti Mariyam
Abdull Hamed, Haza Nuzly
author_facet Saleh, Abdulrazak Yahya
Shamsuddin, Siti Mariyam
Abdull Hamed, Haza Nuzly
author_sort Saleh, Abdulrazak Yahya
title A hybrid differential evolution algorithm for parameter tuning of evolving spiking neural network
title_short A hybrid differential evolution algorithm for parameter tuning of evolving spiking neural network
title_full A hybrid differential evolution algorithm for parameter tuning of evolving spiking neural network
title_fullStr A hybrid differential evolution algorithm for parameter tuning of evolving spiking neural network
title_full_unstemmed A hybrid differential evolution algorithm for parameter tuning of evolving spiking neural network
title_sort hybrid differential evolution algorithm for parameter tuning of evolving spiking neural network
publisher Inderscience Publishers
publishDate 2017
url http://eprints.utm.my/id/eprint/97057/
http://dx.doi.org/10.1504/IJCVR.2017.081231
_version_ 1744353708218515456
score 13.211869

Similar Items